Best of GitMay 2026

  1. 1
    Article
    Avatar of redpandaRedpanda·15w

    Engineering Den: Claude Code skills and hooks for frontend teams

    Redpanda's UX and cloud teams have shipped a shared skills and hooks repository for frontend developers using Claude Code. A single install command sets up guardrails that trigger inside Claude Code and Codex during development, automatically checking type safety, linting, formatting, testing, builds, and catching anti-patterns like TypeScript 'any' usage, missing tests, and accessibility gaps. A lifecycle skill guides Claude Code through a structured flow: clarifying questions, planning, test-driven implementation, PR creation, and report generation that can be posted as a GitHub issue or Jira ticket. The post also mentions SandCastle, a community project that gives each Claude Code instance its own Docker container to autonomously iterate on tickets using the same skills and hooks.

  2. 2
    Article
    Avatar of ullwwn37zsilljprgbshiAvijit Dey·14w

    Run Your Entire Dev Stack With One Click

    Git Switch is a native macOS app built with Tauri, Rust, React, and TypeScript that streamlines multi-repo development workflows. It lets developers manage multiple repositories and services from one place, with features like one-click project group startup, branch-aware auto-restart, built-in terminal, smart Git workflows, and AI-generated commit messages. The app addresses the repetitive task of manually starting and restarting frontend, backend, and worker processes across multiple terminal tabs.

  3. 3
    Article
    Avatar of github_updatesGitHub Changelog·16w

    Enterprise Live Migrations is now in public preview

    GitHub has launched Enterprise Live Migrations (ELM) in public preview, offering a new way to migrate repositories from GitHub Enterprise Server (GHES) to GitHub Enterprise Cloud with data residency. Unlike traditional migrations requiring code freezes, ELM continuously syncs data so developers can keep working throughout the process, with cutover taking minutes rather than days. It is purpose-built for large monorepos with deep git history and high activity. ELM runs as a CLI-driven service on the GHES appliance and is available in GHES versions 3.17.14+, 3.18.8+, 3.19.5+, and 3.20.2+. It complements the existing GitHub Enterprise Importer (GEI) tool and can be run concurrently with it.

  4. 4
    Article
    Avatar of itsfossIt's Foss·16w

    Typical Microsoft! Turns Out VS Code Was Adding Copilot as a Git Co-Author Without Telling Anyone

    VS Code quietly began adding a 'Co-authored-by: Copilot' trailer to git commits after a pull request in April changed the default value of the git.addAICoAuthor setting from 'off' to 'all'. This meant any user who hadn't manually configured the setting had Copilot credited in their git history — even on commits written entirely without AI assistance. The trailer also appeared when Copilot features were explicitly disabled, and was appended after commit finalization with no chance to review it. A VS Code team member acknowledged the mistake on Hacker News, apologized, and a fix reverting the default back to 'off' has been merged and is expected in VS Code 1.119.

  5. 5
    Article
    Avatar of xda-developersXDA Developers·15w

    Zed made sense to me only after I stopped trying to replace VS Code

    A personal take on switching from VS Code to Zed, exploring what makes Zed different rather than treating it as a drop-in replacement. Zed offers native performance advantages by keeping only the UI local while running language servers and terminals remotely. It has built-in Git support, integrated tasks, and debugger support, but takes a simpler project model (one folder per project) compared to VS Code's multi-root workspaces. The transition is eased by importing VS Code settings and enabling a VS Code-style keymap. Key trade-offs include a simpler Git interface lacking advanced tooling like GitLens, and a smaller extension ecosystem overall.

  6. 6
    Article
    Avatar of phProduct Hunt·15w

    Khaos Brain: Local predictive memory for AI agents

    Khaos Brain is a local-first predictive memory system designed for AI agents. It converts task experience, preferences, workflow lessons, and skill-use evidence into Git-versioned memory cards. Agents retrieve relevant cards before starting work and write observations afterward. A Sleep/Dream/Architect maintenance cycle keeps the memory library organized and human-reviewable rather than becoming an opaque black-box store.

  7. 7
    Article
    Avatar of opensourcesquadOpen Source·14w

    Run Your Entire Dev Stack With One Click

    Git Switch is a native macOS app built with Tauri, Rust, React, and TypeScript that streamlines multi-repo development workflows. It lets developers manage multiple repositories and services from one place, with features like one-click project group startup, branch-aware auto-restart, built-in terminal, smart Git workflows, and AI-generated commit messages. The app addresses the repetitive task of manually starting and restarting frontend, backend, and worker processes across multiple terminal tabs when switching branches.

  8. 8
    Article
    Avatar of bytebytegoByteByteGo·15w

    EP214: Claude Code vs. OpenClaw: 5 Design Dimensions

    A newsletter edition comparing Claude Code and OpenClaw across five architectural dimensions: system scope (short-lived vs. long-running daemon), agent runtime (single async loop vs. per-session queues), extension architecture (MCP/hooks vs. manifest-first plugins), memory handling (CLAUDE.md vs. hybrid vector/keyword search), and multi-agent routing (lead-to-subagent vs. route-and-delegate). Additional sections cover AI evaluation frameworks for production apps, how deepfake video pipelines work, and why git revert causes conflicts.

  9. 9
    Article
    Avatar of bradfrostBrad Frost·14w

    Big AI & Design Systems Course updates!

    Brad Frost and team announce that Chapters 1 and 2 of their AI & Design Systems Course are now live, totaling over 16 hours of video content. Chapter 1 covers core concepts including the history and value of design systems and AI, while Chapter 2 focuses on environment setup, AI tooling, API keys, and Git/GitHub workflows. Chapters 3–5 are in production. The course price is increasing from $500 to $995 after May 25th, with purchase power parity options available for students, freelancers, and those in different geographic regions.